The Wall Street Journal reported today that banks spent about $29 billion on consultants last year, and that much of the total was spent on stress tests in the U.S. and abroad. We’re not sure whether either the reported total or the portion spent on stress tests is accurate, but we’re sure it was a LOT of money, and that much of the spending generated no long-term benefit or cost savings. We think this happens for several reasons that we’ll describe below, and we know that with some forethought, it can be avoided. In fact, that’s our value proposition.

The article, Stress Test Inc.: Billions of Dollars, Bank Consultants to Manage Other Consultants, describes banks hiring consultants to manage other consultants and consultants advertising to stop the hiring of consultants–presumably, they mean other consultants.

In general, both the banks and consultants are to blame.

Banks executives tend to overspend for several reasons. These include:

1. Using the sums spent and the number of employees and consultants as signals that they are committed to satisfying supervisory expectations. It is much easier for others to observe (and, therefore, for the banks to signal) these quantity measures rather than quality measures like, say, the comprehensiveness of their risk identification process or, say, the conceptual soundness of a wholesale loss-forecasting model. For a formal, mathematical example of such behavior, see Imprecision in Accounting Measurement: Can It Be Value Enhancing? by Andy and his co-authors, Chandra Kanodia and Raj Singh, in The Journal of Accounting Research (43(3):487-519 May, 2005) or see an earlier version, Optimal Imprecision and Ignorance, at Researchgate.
2. Given the scarceness of resources and the opportunity cost of not having enough, attempting to acquire or reserve as many resources as soon as possible after receiving the CCAR or DFAST review letter, and then developing plans for use.
3. Ignoring (a) priorities, (c) long-term considerations and (c) broader uses. (A) and (b) are related, but if (i) the bank’s priorities coincide with the regulators and (ii) strong progress is made on the biggest problems, then examiners are often forgiving of (or ignore) the little things. Regarding (c), building, say, CCAR or DFAST models without considering alternative uses–either today or in the near future–is simply myopic and ignores one of the easiest ways to generate value from the high costs.
4. Ignoring the root problem(s), which frequently relate to data. Note our maxim: if the data are good enough for modeling, it’s highly likely they are good for every other use, but the converse isn’t true. See The Symptom, not the only Problem and next week, look for our post on information versus data. We have also observed the high cost of having the wrong, non-analytical person in a key role thereby always requiring outside help.

Consultants tend to create inefficiencies, too. Reasons include:

1. Not understanding the problem, objective or constraints, especially when those on site lack detailed banking or institutional knowledge. This can occur when, say, management consultants or accountants attempt to morph into stress-testing or modeling experts. While governance and the process are important, these folks will overemphasize governance, processing and bureaucracy at the expense of understanding the risks, mapping those risks, and mathematically representing those risks and potential losses. A second way is the usual bait-and-switch: the relationship manager is an expert, but he or she doesn’t have time for your firm. The work is being accomplished by inexperienced, recent grads, contract workers or others with no context and no knowledge of the industry, geography, or portfolio. (It’s like having and using a (statistical) recipe without knowledge of the ingredients. Would you want to consume the output?) These folks frequently have less expertise than similarly-experienced associates within the client firm, and often overemphasize (irrelevant) statistical tests, which leads to massive data-mining and the over-fitting of regression equations.
2. Attempting to morph or extrude an existing product or service or model into something that is useful for CCAR/DFAST. We have seen cases where Rube Goldberg, himself, would find the connection or relationship to be convoluted and preposterous. Fortunately, one can usually identify these contraptions by their complication and lack of relevant documentation.

As the article mentions, stress testing is one, very important way to discover flaws and gaps in risk management, databases, operations and other systems. For that reason, we would imagine that a decent portion of the cost associated with stress test consultants, software, and IT services should actually be assigned to pre-existing problems and areas. We suspect they are assigned to stress testing either because the problem was discovered there, or because of the area’s high profile, stress test executives are able to gain resources to fix basic functions, where others couldn’t.

At Spero Risk Associates, we create value by reducing clients’ long-term costs through “thought before calculation” and “zero-defect modeling.” We want to stop your firm’s tragic cycle of throwing it away and redoing it every year. Our goal is to turn your consulting cost into a long-term asset, not an annual expense (or perpetuity). For more information, please call 1-844-SPEROCO or complete our contact form.